PIDAT – PID controller with relay autotuner

Block SymbolLicensing group: AUTOTUNING
PIC

Function Description
The PIDAT block has the same control function as the PIDU block. Additionally it is equipped with the relay autotuning function.

In order to perform the autotuning experiment, it is necessary to drive the system to approximately steady state (at a suitable working point), choose the type of controller to be autotuned (PI or PID) and activate the TUNE input by setting it to on. The controlled process is regulated by special adaptive relay controller in the experiment which follows. One point of frequency response is estimated from the data measured during the experiment. Based on this information the controller parameters are computed. The amplitude of the relay controller (the level of system excitation) and its hysteresis is defined by the amp and hys parameters. In case of hys=0 the hysteresis is determined automatically according to the measurement noise properties on the controlled variable signal. The signal TBSY is set to onduring the tuning experiment. A successful experiment is indicated by and the controller parameters can be found on the outputs pk, pti, ptd, pnd and pb. The c weighting factor is assumed (and recommended) c=0. A failure during the experiment causes TE = on and the output ite provides further information about the problem. It is recommended to increase the amplitude amp in the case of error. The controller is equipped with a built-in function which decreases the amplitude when the deviation of output from the initial steady state exceeds the maxdev limit. The tuning experiment can be prematurely terminated by activating the TBRK input.

Inputs

dv

Feedforward control variable

double

sp

Setpoint variable

double

pv

Process variable

double

tv

Tracking variable

double

hv

Manual value

double

MAN

Manual or automatic mode

bool

off ..

Automatic mode

on ...

Manual mode

TUNE

Start the tuning experiment

bool

TBRK

Stop the tuning experiment

bool

Outputs

mv

Manipulated variable (controller output)

double

de

Deviation error

double

SAT

Saturation flag

bool

off ..

The controller implements a linear control law

on ...

The controller output is saturated

TBSY

Tuner busy flag

bool

TE

Tuning error

bool

off ..

Autotuning successful

on ...

An error occurred during the experiment

ite

Error code; expected time (in seconds) to finishing the tuning experiment while the tuning experiment is active

long

1000 .

Signal/noise ratio too low

1001 .

Hysteresis too high

1002 .

Too tight termination rule

1003 .

Phase out of interval

pk

Proposed controller gain

double

pti

Proposed integral time constant

double

ptd

Proposed derivative time constant

double

pnd

Proposed derivative component filtering

double

pb

Proposed weighting factor – proportional component

double

Parameters

irtype

Controller type (control law)  6

long

1 ....

D

2 ....

I

3 ....

ID

4 ....

P

5 ....

PD

6 ....

PI

7 ....

PID

RACT

Reverse action flag

bool

off ..

Higher mv higher pv

on ...

Higher mv lower pv

k

Controller gain K  1.0

double

ti

Integral time constant Ti  4.0

double

td

Derivative time constant Td  1.0

double

nd

Derivative filtering parameter N  10.0

double

b

Setpoint weighting – proportional part  1.0

double

c

Setpoint weighting – derivative part

double

tt

Tracking time constant. No meaning for controllers without integrator.  1.0

double

hilim

Upper limit of the controller output  1.0

double

lolim

Lower limit of the controller output  -1.0

double

iainf

Type of apriori information  1

long

1 ....

No apriori information

2 ....

Astatic process (process with integration)

3 ....

Low order process

4 ....

Static process + slow closed loop step response

5 ....

Static process + middle fast (normal) closed loop step response

6 ....

Static process + fast closed loop step response

k0

Static gain of the process (must be provided in case of iainf = 3,4,5)  1.0

double

n1

Maximum number of half-periods for estimation of frequency response point  20

long

mm

Maximum number of half-periods for averaging  4

long

amp

Relay controller amplitude  0.1

double

uhys

Relay controller hysteresis

double

ntime

Length of noise amplitude estimation period at the beginning of the tuning experiment [s]  5.0

double

rerrap

Termination value of the oscillation amplitude relative error  0.1

double

aerrph

Termination value of the absolute error in oscillation phase  10.0

double

maxdev

Maximal admissible deviation error from the initial steady state  1.0

double

It is recommended not to change the parameters n1, mm, ntime, rerrap and aerrph.

2019 © REX Controls s.r.o., www.rexygen.com